储能科学与技术 ›› 2025, Vol. 14 ›› Issue (10): 3814-3823.doi: 10.19799/j.cnki.2095-4239.2025.0404

• 储能系统与工程 • 上一篇    下一篇

基于遗传算法的风光储耦合并网制氢系统多目标优化研究

王彦文1(), 魏博1, 张丽芳1, 潘晓君1, 刘如祎1, 郭淼1, 陈康2, 段杨龙2, 叶锋1(), 彭怀午2, 徐超1   

  1. 1.华北电力大学，北京 102206
    2.中国电建集团西北勘测设计研究院有限公司，陕西 西安 710065
  • 收稿日期:2025-04-24 修回日期:2025-05-09 出版日期:2025-10-28 发布日期:2025-10-20
  • 通讯作者: 叶锋 E-mail:15195965577@163.com;fye@ncepu.edu.cn
  • 作者简介:王彦文（1998—），男，硕士研究生，研究方向为可再生能源制氢系统，E-mail：15195965577@163.com；
  • 基金资助:
    国家重点研发计划项目(2021YFB4000100);国家自然科学基金资助项目(22278125)

Multi-objective optimization of the grid-connected wind-solar-storage coupled hydrogen production system based on a genetic algorithm

Yanwen WANG1(), Bo WEI1, Lifang ZHANG1, Xiaojun PAN1, Ruyi LIU1, Miao GUO1, Kang CHEN2, Yanglong DUAN2, Feng YE1(), Huaiwu PENG2, Chao XU1   

  1. 1.North China Electric Power University, Beijing 102206, China
    2.Northwest Engineering Corporation Limited, Power China, Xi'an 710065, Shaanxi, China
  • Received:2025-04-24 Revised:2025-05-09 Online:2025-10-28 Published:2025-10-20
  • Contact: Feng YE E-mail:15195965577@163.com;fye@ncepu.edu.cn

RichHTML

0

PDF

14

摘要：

“双碳”目标下，可再生能源制氢成为推动能源向低碳、清洁、高效转型的重要途径之一。其中，并网制氢存在新能源消纳能力不足、制氢效率与运行经济性不匹配等问题。本研究构建了以最小化平准制氢成本、新能源上网比例及最大化氢气产量、电解槽等效满负荷小时数的多目标优化模型，采用非支配排序遗传算法II(NSGA-II)获得帕累托(Pareto)解集，并基于熵权法-优劣解距离(TOPSIS)法筛选出最优配置方案，同时引入储能系统作为对照研究。研究结果表明，引入储能的并网制氢系统的上网比例为5.78%，氢气产量提高至2899.3 t，下网电量占新能源发电总量的6.49%；而单纯并网制氢系统的新能源上网比例为6.84%，氢气产量为2771.82 t，下网电量占新能源发电总量的11.34%。可以看出，储能系统明显降低了制氢系统的上网比例和电网下电比例，提高了氢气产量。进一步选取春分、夏至、秋分和冬至4个典型日对优化配置方案的日运行特性进行分析，研究结果表明，引入储能系统在有效缓解新能源波动的同时，降低了电解槽负载波动，显著提升了系统的日内运行稳定性与新能源消纳能力。本研究为并网制氢系统的优化配置提供了理论支撑与工程参考价值。

关键词: 并网制氢, 新能源, 储能系统, 多目标优化

Abstract:

Under the "dual carbon" goals, renewable energy-based hydrogen production has become a key pathway to advance the transition toward a low-carbon, clean, and efficient energy system. However, grid-connected hydrogen production faces challenges such as limited renewable energy accommodation and mismatches between hydrogen production efficiency and operational economics. This study develops a multi-objective optimization model to minimize the levelized cost of hydrogen (LCOH) and maximize the grid-feed-in ratio of renewable energy, hydrogen yield, and equivalent full-load hours of the electrolyzer. The Non-dominated Sorting Genetic Algorithm II (NSGA-II) is employed to generate a Pareto solution set, and the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) method with entropy-based weights is applied to identify the optimal configuration. A comparative analysis is performed with and without an energy storage system. Results show that the grid-connected hydrogen production system with energy storage achieves a grid-feed-in ratio of 5.78%, a hydrogen yield of 2899.3 tons, and a curtailment ratio of 6.49% relative to total renewable generation. In contrast, the system without energy storage records a higher grid-feed-in ratio of 6.84%, a lower hydrogen yield of 2771.82 tons, and a curtailment ratio of 11.34%. These findings demonstrate that energy storage effectively reduces grid feed-in and curtailment, thereby enhancing hydrogen production. Furthermore, the operational characteristics of the optimized system on four representative days—vernal equinox, summer solstice, autumnal equinox, and winter solstice are examined. Results indicate that integrating energy storage not only mitigates renewable energy fluctuations but also reduces electrolyzer load variability, thereby improving intra-day operational stability and renewable energy utilization. This study provides theoretical support and engineering reference for the optimal configuration of grid-connected hydrogen production systems.

Key words: grid-connected hydrogen production, new energy, energy storage system, multi-objective optimization

中图分类号: